A novel optical fibre analysis system for particle accelerators

粒子加速器中发生的一些过程，例如从射束中逃逸的杂散粒子（射束损失），可能会损坏或照射机器。这可能会妨碍或阻止加速器的功能，并且辐射延迟机器维护，因为它产生的任何残留放射性必须在工作开始之前衰减到安全水平。粒子加速器用户设施通常旨在最大限度地减少机器停机时间，以向设施用户提供最大的可用射束时间。由于损坏的机器需要停机进行维修[1]，而高辐照水平会延长机器停机时间，因此防止损坏和辐照对这一目标至关重要。通常情况下，这是通过监测束流损失来执行的，以确保其发生保持在安全水平。局部检测器，如电离室（IC），通常用于监测束流损失。然而，它们每个都只能单独监测一个小区域;因此使用IC精确确定光束损失的位置是不可行的。[2]RF加速腔有时会在其中发生火花（RF击穿）。这会损坏腔体表面，也会损坏速调管放大器，因为输入的RF功率会被反射[3]，[4]。传统的RF击穿检测器会跟随伴随击穿的电流或压力的增加[5]。由于这些探测器在击穿发生后才做出响应，因此现有的损伤缓解方法仅限于反应性措施。通过光纤探测器提供了一种解决方案，通过这些事件产生的次级粒子簇射来检测射束损失和RF击穿。这些相对论粒子通过光纤产生切伦科夫光，传播到光电探测器。然后使用飞行时间计算来确定损失位置[6]。这些探测器是非侵入性的，可以提供对机器的连续覆盖;同时监测射束损失和RF故障;快速响应时间（大约几纳秒）;以及大约几厘米的事件位置分辨率（对于当前版本的探测器系统）。与集成电路不同，光纤对磁场和X射线不敏感。这使得光纤可以被放置在许多其他探测器无法放置的区域，例如在像波荡器这样的磁性结构内[7]。光纤技术已经在各种加速器上取得了成功，在这些加速器中，它已经证明了足够的能力来保护单个加速器部分[8，9]甚至整个加速器[10];除了成功地监测暗电流和RF腔中的击穿[4]，[11]。该项目将开发和基准测试新版本的光纤探测器系统，并展示其作为能量回收直线加速器（ERLs）机器保护系统的光束损失和RF击穿分辨率的能力。ERLs是一种新型加速器，旨在通过从使用过的高能光束中提取能量并随后将其转移到新的低能光束中来减少加速器的能量需求。使用过的和新鲜的粒子束（粒子束的子段）通常同时在ERL内循环。当代的例子包括康奈尔大学的CBETA [12]和CERN提出的LHeC对撞机[13]。该应用的关键挑战包括区分光束损失信号与快速重复光束的单个束团，例如通常在ERL中产生的束团;以及区分这些束团的能量以测量每个束团通过ERL的进度。该项目还旨在将机器学习引入oBLM系统[3]，[14]，用于预测和预防机器损坏事件。这使得能够实施主动而不是被动的干预方法。测试将在多个机构进行，如CERA（达雷斯伯里实验室）、CLEAR（欧洲核子研究组织）和CBETA（康奈尔大学）。

Some processes occurring within particle accelerators, such as stray particles escaping from the beam (beam loss), can damage or irradiate the machine. This can impede or prevent the function of the accelerator, and the irradiation delays machine maintenance as any residual radioactivity it creates must decay to safe levels before work can begin.Particle accelerator user facilities commonly aim to minimise machine downtime to offer maximum availability of beam time to facility users. Since a damaged machine requires downtime for repairs [1], and high irradiation levels extend the machine downtime, prevention of both damage and irradiation is important to this goal. Typically, this is performed through monitoring of beam losses to ensure its occurrences are kept to safe levels.Localised detectors such as ionisation chambers (ICs) are commonly used to monitor beam losses. However, they can each individually only monitor a small area; hence it is unfeasible to precisely determine locations of beam loss using ICs. [2] RF accelerating cavities can sometimes undergo sparking within them (RF breakdown). This damages the cavity surface and also the klystron amplifiers as the incoming RF power is reflected [3], [4]. Conventional RF breakdown detectors follow the increase in current or pressure that accompanies a breakdown [5]. Since these detectors respond after the onset of breakdown, the available damage mitigation methods are limited to reactive measures.A solution is offered through optical fibre-based detectors to detect beam losses and RF breakdowns through the showers of secondary particles these events produce. Passage of these relativistic particles through an optical fibre produces Cherenkov light, which propagates to photodetectors. A time-of-flight calculation is then used to determine the loss location [6]. These detectors are non-invasive which can offer, continuous coverage of a machine; simultaneous monitoring of both beam losses and RF breakdowns; a fast response time (on the order of a few nanoseconds); and event location resolution on the order of a few cm (for the current version of the detector system). Unlike ICs, optical fibres are insensitive to magnetic fields and X-rays. This allows fibres to be positioned in areas where many other detectors cannot, such as inside magnetic structures like undulators [7].Optical fibre technology has been successful on a variety of accelerators, where it has demonstrated sufficient capability to protect individual accelerator sections [8, 9] or even entire accelerators [10]; in addition to successfully monitoring dark current and breakdowns in RF cavities [4], [11]. This project will develop and benchmark the new version of the optical fibre detector system and demonstrate its capabilities for beam loss and RF breakdown resolution as a machine protection system for energy recovery linear accelerators (ERLs).ERLs are a type of novel accelerator that seek to decrease accelerator energy demands by extracting energy from its used, high-energy beams and subsequently transferring it to fresh, low-energy beams. Both used and fresh particle bunches (sub-segments of beams) often circulate simultaneously within the ERL. Contemporary examples include CBETA at Cornell University [12], and the LHeC collider proposed by CERN [13]. Key challenges for this application include distinguishing beam loss signals from individual bunches of fast-repeating beams, such as those typically produced in ERLs; and discriminating between the energies of these bunches to gauge the progress of each bunch through the ERL.This project also aims to introduce machine learning to the oBLM system [3], [14], for the prediction and prevention of machine damaging events. This allows for the implementation of proactive rather than reactive intervention methods. Testing will be carried out at various facilities, such as CLARA (Daresbury Laboratory), CLEAR (CERN), and CBETA (Cornell).